Datasets:

Multilinguality:
multilingual
Size Categories:
1K<n<10K
Annotations Creators:
crowdsourced
ArXiv:
Tags:
License:
albertvillanova HF staff commited on
Commit
33f040f
1 Parent(s): f9d9a27

Add X-CSQA-sw data files

Browse files
README.md CHANGED
@@ -844,13 +844,13 @@ dataset_info:
844
  dtype: string
845
  splits:
846
  - name: test
847
- num_bytes: 222517
848
  num_examples: 1074
849
  - name: validation
850
- num_bytes: 211708
851
  num_examples: 1000
852
- download_size: 7519903
853
- dataset_size: 434225
854
  - config_name: X-CSQA-ur
855
  features:
856
  - name: id
@@ -1005,6 +1005,12 @@ configs:
1005
  path: X-CSQA-ru/test-*
1006
  - split: validation
1007
  path: X-CSQA-ru/validation-*
 
 
 
 
 
 
1008
  - config_name: X-CSQA-vi
1009
  data_files:
1010
  - split: test
 
844
  dtype: string
845
  splits:
846
  - name: test
847
+ num_bytes: 222215
848
  num_examples: 1074
849
  - name: validation
850
+ num_bytes: 211426
851
  num_examples: 1000
852
+ download_size: 214954
853
+ dataset_size: 433641
854
  - config_name: X-CSQA-ur
855
  features:
856
  - name: id
 
1005
  path: X-CSQA-ru/test-*
1006
  - split: validation
1007
  path: X-CSQA-ru/validation-*
1008
+ - config_name: X-CSQA-sw
1009
+ data_files:
1010
+ - split: test
1011
+ path: X-CSQA-sw/test-*
1012
+ - split: validation
1013
+ path: X-CSQA-sw/validation-*
1014
  - config_name: X-CSQA-vi
1015
  data_files:
1016
  - split: test
X-CSQA-sw/test-00000-of-00001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:801f1a8e405bd3362921c1e00450e26ca020cfbe8ef3506a6c743458ab1e0f14
3
+ size 111234
X-CSQA-sw/validation-00000-of-00001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4d6c93bddbdb4fe71535d4f6de587e240103f3e86309690b9bac7243be287bba
3
+ size 103720
dataset_infos.json CHANGED
@@ -931,48 +931,38 @@
931
  "features": {
932
  "id": {
933
  "dtype": "string",
934
- "id": null,
935
  "_type": "Value"
936
  },
937
  "lang": {
938
  "dtype": "string",
939
- "id": null,
940
  "_type": "Value"
941
  },
942
  "question": {
943
  "stem": {
944
  "dtype": "string",
945
- "id": null,
946
  "_type": "Value"
947
  },
948
  "choices": {
949
  "feature": {
950
  "label": {
951
  "dtype": "string",
952
- "id": null,
953
  "_type": "Value"
954
  },
955
  "text": {
956
  "dtype": "string",
957
- "id": null,
958
  "_type": "Value"
959
  }
960
  },
961
- "length": -1,
962
- "id": null,
963
  "_type": "Sequence"
964
  }
965
  },
966
  "answerKey": {
967
  "dtype": "string",
968
- "id": null,
969
  "_type": "Value"
970
  }
971
  },
972
- "post_processed": null,
973
- "supervised_keys": null,
974
- "task_templates": null,
975
  "builder_name": "xcsr",
 
976
  "config_name": "X-CSQA-sw",
977
  "version": {
978
  "version_str": "1.1.0",
@@ -984,27 +974,20 @@
984
  "splits": {
985
  "test": {
986
  "name": "test",
987
- "num_bytes": 222517,
988
  "num_examples": 1074,
989
- "dataset_name": "xcsr"
990
  },
991
  "validation": {
992
  "name": "validation",
993
- "num_bytes": 211708,
994
  "num_examples": 1000,
995
- "dataset_name": "xcsr"
996
- }
997
- },
998
- "download_checksums": {
999
- "https://inklab.usc.edu/XCSR/xcsr_datasets.zip": {
1000
- "num_bytes": 7519903,
1001
- "checksum": "c45b29ece740643252d5402e76be1e33f96f9d6910053f79e80d39887f10c85e"
1002
  }
1003
  },
1004
- "download_size": 7519903,
1005
- "post_processing_size": null,
1006
- "dataset_size": 434225,
1007
- "size_in_bytes": 7954128
1008
  },
1009
  "X-CSQA-ur": {
1010
  "description": "To evaluate multi-lingual language models (ML-LMs) for commonsense reasoning in a cross-lingual zero-shot transfer setting (X-CSR), i.e., training in English and test in other languages, we create two benchmark datasets, namely X-CSQA and X-CODAH. Specifically, we automatically translate the original CSQA and CODAH datasets, which only have English versions, to 15 other languages, forming development and test sets for studying X-CSR. As our goal is to evaluate different ML-LMs in a unified evaluation protocol for X-CSR, we argue that such translated examples, although might contain noise, can serve as a starting benchmark for us to obtain meaningful analysis, before more human-translated datasets will be available in the future.\n",
 
931
  "features": {
932
  "id": {
933
  "dtype": "string",
 
934
  "_type": "Value"
935
  },
936
  "lang": {
937
  "dtype": "string",
 
938
  "_type": "Value"
939
  },
940
  "question": {
941
  "stem": {
942
  "dtype": "string",
 
943
  "_type": "Value"
944
  },
945
  "choices": {
946
  "feature": {
947
  "label": {
948
  "dtype": "string",
 
949
  "_type": "Value"
950
  },
951
  "text": {
952
  "dtype": "string",
 
953
  "_type": "Value"
954
  }
955
  },
 
 
956
  "_type": "Sequence"
957
  }
958
  },
959
  "answerKey": {
960
  "dtype": "string",
 
961
  "_type": "Value"
962
  }
963
  },
 
 
 
964
  "builder_name": "xcsr",
965
+ "dataset_name": "xcsr",
966
  "config_name": "X-CSQA-sw",
967
  "version": {
968
  "version_str": "1.1.0",
 
974
  "splits": {
975
  "test": {
976
  "name": "test",
977
+ "num_bytes": 222215,
978
  "num_examples": 1074,
979
+ "dataset_name": null
980
  },
981
  "validation": {
982
  "name": "validation",
983
+ "num_bytes": 211426,
984
  "num_examples": 1000,
985
+ "dataset_name": null
 
 
 
 
 
 
986
  }
987
  },
988
+ "download_size": 214954,
989
+ "dataset_size": 433641,
990
+ "size_in_bytes": 648595
 
991
  },
992
  "X-CSQA-ur": {
993
  "description": "To evaluate multi-lingual language models (ML-LMs) for commonsense reasoning in a cross-lingual zero-shot transfer setting (X-CSR), i.e., training in English and test in other languages, we create two benchmark datasets, namely X-CSQA and X-CODAH. Specifically, we automatically translate the original CSQA and CODAH datasets, which only have English versions, to 15 other languages, forming development and test sets for studying X-CSR. As our goal is to evaluate different ML-LMs in a unified evaluation protocol for X-CSR, we argue that such translated examples, although might contain noise, can serve as a starting benchmark for us to obtain meaningful analysis, before more human-translated datasets will be available in the future.\n",